New method of fault diagnosis of rotating machinery based on distance of information entropy

Abstract

This paper introduces the basic conception of information fusion and some fusion diagnosis methods commonly used nowadays in rotating machinery. From the thought of the information fusion, a new quantitative feature index monitoring and diagnosing the vibration fault of rotating machinery, which is called distance of information entropy, is put forward on the basis of the singular spectrum entropy in time domain, power spectrum entropy in frequency domain, wavelet energy spectrum entropy, and wavelet space feature entropy in time-frequency domain. The mathematic deduction suggests that the conception of distance of information entropy is accordant with the maximum subordination principle in the fuzzy theory. Through calculation it has been proved that this method can effectively distinguish different fault types. Then, the accuracy of rotor fault diagnosis can be improved through the curve chart of the distance of information entropy at multi-speed.

Key words： rotating machinery; information fusion; fault diagnosis; Information entropy; distance of the information entropy

References

1 Yuan X H, Qu L S. Utility of information fusion in mechanical fault diagnosis. Journal of Vibration Measurement & Diagnosis, 1999, 19(3): 187-192

2 Cai X G, Ma P. Study on simultaneous fault diagnosis based information fusion technique. In: Proceedings of the CSEE, 2003, 23(5): 112-115

3 Ling W Y, Jia M P, Xu F Y. Optimizing strategy on rough set neural network fault diagnosis system. In:Proceedings of the CSEE, 2003, 23(5): 98-102

4 Yu B S, Huang W H. A new model of self-organized neural network for extraction diagnostic rules. In: Proceedings of the CSEE, 2001, 21(3): 16-18, 61

5 Shen T, Huang S H, Han S M, Yang S Z. Extracting information entropy features for rotating machinery vibration signals. Chinese Journal of Mechanical Engineering, 2001, 37(6): 94-98